A wealth of research in demography and social epidemiology has found that two of the most important environmental determinants of health for the elderly are socioeconomic status and social relationships. One relatively unexplored pathway linking these upstream factors with health outcomes is sleep. Recent studies have found that socioeconomic factors are correlated with sleep, and there is mounting evidence that connects sleep and health. Few population- based data are available that would allow researchers to test whether social and economic factors affect health in part through their impact on sleep, particularly with objective measures of sleep. The National Social Life, Health and Aging Project (NSHAP), a population-based, longitudinal study of health and social processes among older adults in the U.S., obtained extensive questionnaire and biomeasure data from a national probability sample of 3,005 community- dwelling adults ages 57-85 in 2005/6. The same individuals are being reinterviewed five years later, and the cohort expanded by recruiting spouses. Among the strengths of NSHAP are detailed marital quality and social network data, and novel home-based biomeasure collection. During the 5-year follow-up interval there will have been changes due to marital loss, declines in partner health, similar losses in the social network, and changed economic circumstances, many due to the recession. As part of a substudy that was not in the original NSHAP proposal, sleep data are being collected on one-third of the cohort and their spouses using three nights of wrist actigraphy. (projected sleep substudy sample 788). We propose using this unique data resource to develop and test models of the effects of social and economic circumstances and 5- year changes in these circumstances on specific psychosocial and physiologic domains, through their impact on sleep. We hypothesize that social and economic factors influence health in part through their effects on sleep behavior, quality and quantity. We will examine how demographic factors, socioeconomic status, and social relationships relate to concurrently measured sleep behavior (time set aside for sleep and bedtime) and actigraph-measured sleep characteristics (duration and several measures of quality) in older adults, and we will explore how the sleep and health of one spouse influence the other spouse's sleep. We will examine how 5-year changes in social or economic circumstances predict changes in psychosocial characteristics (i.e., loneliness, anxiety, depression) or predict indicators of physiological function (cognition, inflammatory markers, cortisol stress response) and the extent to which those effects may be mediated by sleep behavior or the quantity or quality of measured sleep.